Overview

The RJTPP/scot0402s-deepseek-14b-full is a 14.8 billion parameter language model developed by RJTPP. It is finetuned from the unsloth/DeepSeek-R1-Distill-Qwen-14B-unsloth-bnb-4bit base model, indicating its foundation in the Qwen2 architecture. This model leverages Unsloth and Huggingface's TRL library for its training process.

Key Characteristics

• Architecture: Based on the Qwen2 model family.
• Parameter Count: 14.8 billion parameters.
• Training Efficiency: Achieved 2x faster training by utilizing Unsloth and Huggingface's TRL library.
• License: Released under the Apache-2.0 license.

Potential Use Cases

This model is suitable for a variety of general language understanding and generation tasks, benefiting from its efficient finetuning process. Its foundation in the DeepSeek-R1-Distill-Qwen architecture suggests capabilities in areas such as:

• Text generation
• Question answering
• Summarization
• Code-related tasks (given its DeepSeek lineage, though not explicitly stated for this finetune)